Ship with confidence
Simple, usage-based pricing. No surprises.
4M tokens included
8M tokens included
Core Features
Everything you need to observe and understand your LLM applications.
- · SDK tracing with automatic span extraction
- · Local, S3, or Postgres storage
- · CLI and TUI for exploring traces
- · Diagnostics for system health checks
- · Analyze for cost, quality, and latency insights
- · Assertions for automated quality checks
- · Human grading via keyboard-driven TUI
AI-Powered Features
LLM-powered capabilities that automate evaluation and improvement.
- · judge — automated quality scoring at scale
- · improve — offline testing before deploying
- · AI Engineer Loop — autonomous observe → judge → analyze → recommend
- · Briefings — daily or weekly performance reports delivered to email or Slack
These features consume tokens from your monthly allocation.
Frequently asked questions
What's included in the free trial?
30 days of full Team access with 100k tokens. No credit card required. Try judge runs, calibrate rubrics, and explore all AI-powered features before upgrading.
How do tokens work?
AI-Powered features make LLM calls on your behalf. Each call consumes tokens. Your plan includes a set amount per month. Overage is billed at $10/million tokens.
What counts as a token?
Every LLM call Bandito makes. A typical judge run against 100 traces uses ~5 million tokens depending on trace complexity.
Can I upgrade anytime?
Yes. Upgrades take effect immediately. Downgrades apply at the end of your billing cycle.
What happens if I go over my token limit?
We'll notify you at 80% and 100%. You can continue using the service and we'll bill overage at $10/million (Team) or $8/million (Enterprise).
Ready to ship with confidence?
Join the waitlist to get early access.